Merge pull request #74 from causy-dev/pc-classic

feat(causal_discovery_algorithms): add PC classic without runtime opt…
causy-dev · Jan 29, 2025 · 0dc0f33 · 0dc0f33
2 parents 743570d + 229de5d
commit 0dc0f33
Show file tree

Hide file tree

Showing 2 changed files with 35 additions and 65 deletions.
diff --git a/causy/causal_discovery/constraint/algorithms/pc.py b/causy/causal_discovery/constraint/algorithms/pc.py
@@ -88,6 +88,38 @@
     )
 )
 
+PCClassic = graph_model_factory(
+    Algorithm(
+        pipeline_steps=[
+            CalculatePearsonCorrelations(
+                display_name="Calculate Pearson Correlations"
+            ),
+            CorrelationCoefficientTest(
+                threshold=VariableReference(name="threshold"),
+                display_name="Correlation Coefficient Test",
+            ),
+            ExtendedPartialCorrelationTestMatrix(
+                threshold=VariableReference(name="threshold"),
+                display_name="Extended Partial Correlation Test Matrix",
+                generator=PairsWithNeighboursGenerator(
+                    comparison_settings=ComparisonSettings(
+                        min=3, max=AS_MANY_AS_FIELDS
+                    ),
+                    shuffle_combinations=False,
+                ),
+            ),
+            *PC_ORIENTATION_RULES,
+            ComputeDirectEffectsInDAGsMultivariateRegression(
+                display_name="Compute Direct Effects in DAGs Multivariate Regression"
+            ),
+        ],
+        edge_types=PC_EDGE_TYPES,
+        extensions=[PC_GRAPH_UI_EXTENSION],
+        name="PC",
+        variables=[FloatVariable(name="threshold", value=PC_DEFAULT_THRESHOLD)],
+    )
+)
+
 PCStable = graph_model_factory(
     Algorithm(
         pipeline_steps=[

diff --git a/tests/test_pc_e2e.py b/tests/test_pc_e2e.py
@@ -3,7 +3,7 @@
     PC,
     PC_ORIENTATION_RULES,
     PC_GRAPH_UI_EXTENSION,
-    PC_DEFAULT_THRESHOLD,
+    PC_DEFAULT_THRESHOLD, PCClassic,
 )
 from causy.causal_effect_estimation.multivariate_regression import (
     ComputeDirectEffectsInDAGsMultivariateRegression,
@@ -351,37 +351,6 @@ def test_tracking_triples_four_nodes(self):
         self.assertEqual(len(triples), 6 + 12 + 12)
 
     def test_track_triples_three_nodes_custom_pc(self):
-        algo = graph_model_factory(
-            Algorithm(
-                pipeline_steps=[
-                    CalculatePearsonCorrelations(
-                        display_name="Calculate Pearson Correlations"
-                    ),
-                    CorrelationCoefficientTest(
-                        threshold=VariableReference(name="threshold"),
-                        display_name="Correlation Coefficient Test",
-                    ),
-                    ExtendedPartialCorrelationTestMatrix(
-                        threshold=VariableReference(name="threshold"),
-                        display_name="Extended Partial Correlation Test Matrix",
-                        generator=PairsWithNeighboursGenerator(
-                            comparison_settings=ComparisonSettings(
-                                min=3, max=AS_MANY_AS_FIELDS
-                            ),
-                            shuffle_combinations=False,
-                        ),
-                    ),
-                    *PC_ORIENTATION_RULES,
-                    ComputeDirectEffectsInDAGsMultivariateRegression(
-                        display_name="Compute Direct Effects in DAGs Multivariate Regression"
-                    ),
-                ],
-                edge_types=PC_EDGE_TYPES,
-                extensions=[PC_GRAPH_UI_EXTENSION],
-                name="PC",
-                variables=[FloatVariable(name="threshold", value=PC_DEFAULT_THRESHOLD)],
-            )
-        )
         rdnv = self.seeded_random.normalvariate
         sample_generator = IIDSampleGenerator(
             edges=[
@@ -391,7 +360,7 @@ def test_track_triples_three_nodes_custom_pc(self):
             random=lambda: rdnv(0, 1),
         )
         test_data, graph = sample_generator.generate(10000)
-        tst = algo()
+        tst = PCClassic()
         tst.create_graph_from_data(test_data)
         tst.create_all_possible_edges()
         pc_results = tst.execute_pipeline_steps()
@@ -407,37 +376,6 @@ def test_track_triples_three_nodes_custom_pc(self):
         self.assertIn(len(triples), [6, 7, 8])
 
     def test_track_triples_two_nodes_custom_pc_unconditionally_independent(self):
-        algo = graph_model_factory(
-            Algorithm(
-                pipeline_steps=[
-                    CalculatePearsonCorrelations(
-                        display_name="Calculate Pearson Correlations"
-                    ),
-                    CorrelationCoefficientTest(
-                        threshold=VariableReference(name="threshold"),
-                        display_name="Correlation Coefficient Test",
-                    ),
-                    ExtendedPartialCorrelationTestMatrix(
-                        threshold=VariableReference(name="threshold"),
-                        display_name="Extended Partial Correlation Test Matrix",
-                        generator=PairsWithNeighboursGenerator(
-                            comparison_settings=ComparisonSettings(
-                                min=3, max=AS_MANY_AS_FIELDS
-                            ),
-                            shuffle_combinations=False,
-                        ),
-                    ),
-                    *PC_ORIENTATION_RULES,
-                    ComputeDirectEffectsInDAGsMultivariateRegression(
-                        display_name="Compute Direct Effects in DAGs Multivariate Regression"
-                    ),
-                ],
-                edge_types=PC_EDGE_TYPES,
-                extensions=[PC_GRAPH_UI_EXTENSION],
-                name="PC",
-                variables=[FloatVariable(name="threshold", value=PC_DEFAULT_THRESHOLD)],
-            )
-        )
         rdnv = self.seeded_random.normalvariate
         sample_generator = IIDSampleGenerator(
             edges=[
@@ -447,7 +385,7 @@ def test_track_triples_two_nodes_custom_pc_unconditionally_independent(self):
             random=lambda: rdnv(0, 1),
         )
         test_data, graph = sample_generator.generate(10000)
-        tst = algo()
+        tst = PCClassic()
         tst.create_graph_from_data(test_data)
         tst.create_all_possible_edges()
         pc_results = tst.execute_pipeline_steps()